mix_\(spring\)springmix_\(snowflake\)snowflakemix_\(shark\)mix_\(sea slug\)sea slugmix_\(preal\)prealmix_\(octopus\)mix_\(jellyfish\)mix_\(frog\)greentransparentmix_\(flower\)mix_\(creature\)graymix_\(coral\)mix_\(bug\)mix_\(sea blue dragon\)mix_\(axolotl\)mix_\(gost\)gost
License Scope:
Creative License Scope
Online Image Generation
Merge
Allow Downloads
Commercial License Scope
Sale or Commercial Use of Generated Images
Resale of Models or Their Sale After Merging
Model Parameters:
Base Model:
SD 1.5
Epochs:
0
Iteration Steps:
0
Clip Skip:
0
Review:
0
Review
Using colors as trigger words in the 1.0 version doesn't seem like a good idea, so version 2.0 uses other trigger words and increases the training steps.
Folder 10_mix_gost: 66 images found
Folder 10_mix_gost: 660 steps
Folder 25_gost: 5 images found
Folder 25_gost: 125 steps
Folder 25_mix_axolotl: 30 images found
Folder 25_mix_axolotl: 750 steps
Folder 25_mix_blueD: 30 images found
Folder 25_mix_blueD: 750 steps
Folder 25_mix_bug: 31 images found
Folder 25_mix_bug: 775 steps
Folder 25_mix_coral: 29 images found
Folder 25_mix_coral: 725 steps
Folder 25_mix_creature: 30 images found
Folder 25_mix_creature: 750 steps
Folder 25_mix_flower: 31 images found
Folder 25_mix_flower: 775 steps
Folder 25_mix_frog: 30 images found
Folder 25_mix_frog: 750 steps
Folder 25_mix_J: 30 images found
Folder 25_mix_J: 750 steps
Folder 25_mix_octopus: 30 images found
Folder 25_mix_octopus: 750 steps
Folder 25_mix_preal: 30 images found
Folder 25_mix_preal: 750 steps
Folder 25_mix_sea slug: 30 images found
Folder 25_mix_sea slug: 750 steps
Folder 25_mix_shark: 29 images found
Folder 25_mix_shark: 725 steps
Folder 25_mix_snow: 30 images found
Folder 25_mix_snow: 750 steps
Folder 25_mix_spring: 31 images found
Folder 25_mix_spring: 775 steps
max_train_steps = 28275
stop_text_encoder_training = 0
lr_warmup_steps = 0
Here is the trainer command as a reference. It will not be executed:
accelerate launch --num_cpu_threads_per_process=2 "train_network.py"
--pretrained_model_name_or_path="D:/ai/stable-diffusion-webui/models/Stable-diffusion/v1-5-pruned.safetensors"
--train_data_dir="D:/ai/test/image"
--resolution=768,768 --output_dir="D:/ai/test/model"
--logging_dir="D:/ai/test/log"
--network_alpha="128"
--save_model_as=safetensors
--network_module=networks.lora
--text_encoder_lr=5e-5
--unet_lr=0.0001
--network_dim=128
--output_name="25_preal_1v5_768_0001BD5e_NR128NA128off"
--lr_scheduler_num_cycles="5"
--learning_rate="0.0001"
--lr_scheduler="constant"
--train_batch_size="2"
--max_train_steps="28275"
--save_every_n_epochs="1"
--mixed_precision="bf16"
--save_precision="bf16"
--seed="1234"
--caption_extension=".txt"
--cache_latents
--optimizer_type="AdamW8bit"
--max_data_loader_n_workers="1"
--clip_skip=2
--bucket_reso_steps=64
--xformers
--bucket_no_upscale
Using colors as trigger words in the 1.0 version doesn't seem like a good idea, so version 2.0 uses other trigger words and increases the training steps.